Thanks so much for your kind and enthusiastic and extensive help.
There are too many responses to reply them all in private.
But I now feel encouraged to wade into the deep.
This week I have to make up my mind about which Linux version.
I will try to install ( at least .. begin with... ) next weekend.
Fortunately, I got so many substantially valuable advices.
Again, I am grateful for that.
I'll keep you informed.
I do hope to get further support in this adventure.
Two more questions :
A bit OT but I should know :
Under which Linux version(s) does, just for testing, the ordinary SoundBlasterLive ! ( not the digital, not the 5.1 ) sound card work ?
Or is it version independent but driver dependent ?
Or depending or something else ?
Or doesn't this give a problem at all ?
I have reasonable NT4/5 knowledge and experience, but not the slightest idea on Linux ( not yet... ).
I'll buy a 2 x 96 Khz card ( to get the best out of Linrad as I seemed to understand ) but only after things are running properly here ever.
I run a P4 1.4 / 768 PC133 : enough for the ultimate out of Linrad ?
I do not need a 96 Khz for other apps, just for Linrad.
I am building the Brooks Shera GPS disciplined frequency standard ( < 10^-11 for a 100 sec. dot ),
HP 10811 OCXO : phase noise : offset 1 Hz < -100dBc.. ....1 KHz <-160 dBc.
I started this frequency standard project for quite different reasons than Linrad.
However, does it make sense to use this standard for LO ( 144 MHz ) in a DC front end for Linrad ?
Or is this overkill ( e.g. : doesn't the SBL-1 or TUF spoil alot as to converter noise c.a. ? ).
I intend to experiment with EME QRSSS ( if another nerd is also interested ... ) and I have the ( naive ? ) idea that short term stability ( for very small FFT bin bandwidth during the dot length ) and oscillator noise cannot be good enough.
Right ? Or overdone ?
Thanks for help.